8 thoughts on “

  1. @beko Same! I’ve been experimenting with NGINX microcaching and a couple overall NGINX and PHP tweaks. It’ll occasionally feel slow, still, but no more crashes. (Or if it does, it comes back up on its own.) I’m still mostly preloading the front-end cache, so that isn’t as much of an issue. The problem is my (“ActivityPub”) posts being boosted or replied to, and it mostly affects uncached pages, so the impact to anonymous visitors is small-ish (I hope).

  2. @beko Hehe, just boosted your post, then clicked the link. Took maybe like 5–10 seconds to load (which is exactly the kind of thing I’m talking about, or at least I think it is) 🤓

  3. Heh yeah 😀 TBF I’m not kidding when I write: This is a backyard hosted website on 2nd hand hardware, which means it sometimes goes offline. Please hold the line 🤓

  4. Think it’s more the amount. There’s more to this than just pushing the note. Federated servers also check version and actor and some other things before pre-fetching images and content. This is quickly enough to saturate a small server on a land-line on 2nd hand hardware 😉

    Didn’t have this sort of trouble when hosting in a data centre but this is simply too expensive nowadays and really enough for the “regular” visitors.

  5. @bekopharm Yeah, weird. I searched for it just now and it popped up on my instance that way. 🙃 But the plugin doesn’t reschedule failed jobs, so if, e.g., WP-Cron times out or the POST request fails because the receiving server is temporarily down, well … 🤷 @beko

Leave a Reply